FPGA implementation of canny edge detection algorithm for X-ray images

Abstract

In the domain of medical imaging, X-ray imaging is the most frequent and commonly used technique for diagnosing diseases of various types. So, quick and exact diagnosis can be critical to the success of any given treatment. For this purpose a hardware accelerator for Canny edge detection algorithm is presented in this work with its implementation on Spartan (3E-XC3S1600) kit and programmed using VHDL. To simplify the hardware implementation some assumptions are made on the traditional algorithm like the intermediate values of Canny operator are calculated with integer values instead of floating point or fixed point values so that the available resources of (Spartan 3E-XC3S1600) platform are sufficient for this purpose in terms of the use of BLOCK RAMs. Also a modified 18-bit VGA port is built for the purpose of displaying the original and result images clearly and accurately instead of the original 3-bit VGA port available on the used Spartan3E kits. A speedup of about were achieved with the proposed design over the software implementation using Mat-Lab on the art personal computer with a maximum area occupation 100%. Although Canny operator produces best results when compared with other edge detection operators because Canny yields thin lines for its edges by using non-maximal suppression, but this is influence on the silicon area occupation.